Sensing and Forecasting of Pollution Data in Mexico City

Abstract

In this paper we present the characteristics of sensors used to monitor the pollution levels in Mexico City, namely sulfur dioxide (SO2), nitrogen oxides (NOx), ozone (O3), , and carbon monoxide (CO). A novel algorithm to predict contamination levels is presented: the Gamma classifier. Also, a new coding technique is introduced, allowing the conversion from a series of values taken from SIMAT databases into a set of patterns, which in turn are useful for the task of pollutant forecasting. Experimental results show a competitive performance by the Gamma classifier as a predictor, when compared to other methods.